Cross-Domain Reuse of Extracted Knowledge in Genetic Programming for Image Classification

Abstract

Genetic programming (GP) is a well-known evolutionary computation technique, which has been successfully used to solve various problems, such as optimisation, image analysis and classification. Transfer learning is a type of machine learning approach that can be used to solve complex tasks. Transfer learning has been introduced to genetic programming to solve complex Boolean and symbolic regression problems with some promise. However, the use of transfer learning with genetic programming has not been investigated to address complex image classification tasks with noise and rotations, where GP cannot achieve satisfactory performance, but GP with transfer learning may improve the performance. In this paper, we propose a novel approach based on transfer learning and genetic programming to solve complex image classification problems by extracting and reusing blocks of knowledge/information, which are automatically discovered from similar as well as different image classification tasks during the evolutionary process. The proposed approach is evaluated on three texture data sets and three office data sets of image classification benchmarks, and achieves better classification performance than the state-of-the-art image classification algorithm. Further analysis on the evolved solutions/trees shows that the proposed approach with transfer learning can successfully discover and reuse knowledge/information extracted from similar or different problems to improve its performance on complex image classification problems.

Transfer Learning-Based Dynamic Multiobjective Optimization Algorithms

“…Recently, there are some works exploiting knowledge reuse techniques or machine learning in evolutionary computation that have been proposed. In [26], the authors propose an approach based on transfer learning and genetic programming to solve complex image classification problems. The basic idea of the proposed algorithm is that the knowledge learned from a simpler subtask is used to solve a more complex subtask, and reusing knowledge blocks are discovered from similar as well as different image classification tasks during the evolutionary process.…”

Artificial Bee Colony Programming Descriptor for Multi-Class Texture Classification

“…The method separated the image into the sub-regions, extracted the Speeded up Robust Features (SURF) points and achieved successful classification results using the Support Vector Machine (SVM) classifier. Iqbal et al proposed an image classification method using transfer learning and GP [8]. The GP trees are extracted by the fragments of the transfer learning improved classification performance producing more useful initial population.…”

“…Adapted version of the distance function according to the logical function (Equation 7) is shown in Equation (8). D b (between distances) calculates the average distance between classes, and D w (within Distances) calculates the average distance within classes.…”

Transfer optimization in complex engineering design

“…[80,114]. Their proposition was demonstrated on boolean problems [80] and classification problems [114] resulting in promising outcomes. Exploring along a different path, in [85], the paradigm of evolutionary multi-tasking was introduced providing the means to simultaneously address multiple tasks (that are related) using a single population of evolving solutions.…”

“…In a relatively recent effort, the reuse of "building-blocks" of knowledge from source tasks for genetic programming (in this case, code fragments) was proposed by Iqbal et al. [80,114]. Their proposition was demonstrated on boolean problems [80] and classification problems [114] resulting in promising outcomes.…”